[postgis-devel] Oracle SDO_GEOMETRY vs PostGIS WKT

[Jorge Arévalo]

Hi,

On Wed, Aug 25, 2010 at 1:08 AM, Martin Davis <mbdavis at refractions.net> wrote:
> Here's my take:
>
> Yes, reference [2] is the best reference for the true SFS Geometry standard.
>  It's an abstract standard, however.  Each implementation has its own
> idiosyncracies, which can only be understood by reading the implementation
> manuals (and more than likely working with the implementation directly).
>

Thanks. A lot of reading and hard work to do.


> It's necessary to distinguish between the naming conventions used by
> vendors, and their underlying implementations.  The naming conventions
> reflect an attempt to (more or less) fulfull the requirements of the SFS and
> ISO standards.  But each vendor has a different implementation, which seeks
> to meet the varying marketing goals of the vendor.
>

I like PostGIS convention, with ST_ prefix.


> Oracle's SDO_GEOMETRY type is more full-featured than is required by SFS,
> which is one reason for its complexity.  Also, it uses existing features of
> the Oracle Object-Relational model, which also means that complexity is
> perhaps more exposed than is desirable.
>

Ok, then I confirm: Oracle goes beyond the standard, and mix it with
its own object model.


> ESRI recently released their own ST_GEOMETRY type for Oracle, presumably in
> an attempt to both better match the requirements of their other software
> products, as well as having the pleasant side effect of "plugging the leak"
> of spatial data from their walled garden to Oracle's more accessible native
> spatial format.
>
> AFAICS, the Oracle ST_GEOMETRY types are provided in order to increase
> compliance with the SQL/MM standard.
>

Good idea, I think, after reading the complex format used for SDO_GEOMETRY.


> It's nice to hear that you find the PostGIS ST_GEOMETRY implementation
> easier to work with and more standards compliant.  I think this is the
> result of a phenomenon which occurs in many open-source projects. Having no
> commercial agenda and limited resources means that easiest path is to simply
> follow a standard as closely as possible, and to remove as many rough edges
> as possible.
>

Yes, I find it simply easier to understand and to use. And fully agree
with you. That's Debian's philosohpy. I like it.


Jorge

> Martin
>
> Jorge Arévalo wrote:
>>
>> Hello,
>>
>> I'm studying Oracle SDO_GEOMETRY type [1], and I find it really
>> tangled. But, AFAIK, the SFS standard [2] ST_Geometry is implemented
>> too, provided by ESRI [3]. And even a third ST_Geometry object,
>> provided by Oracle and basically equal to SDO_GEOMETRY type, is
>> available [4].
>>
>> On the other hand, and please correct me if I'm wrong, PostGIS simply
>> implements the standard ST_Geometry type. I'm reading the SFS
>> documents [2], and the PostGIS documentation [5], and I find it
>> simpler to specify if the geometry is compound or not, if it has
>> holes... this is: the geometry's properties.
>>
>> So, my questions, as a beginner, are:
>> - Why does Oracle use a tangled format like SDO_GEOMETRY, if they
>> could simply implement the easier standard? Any obvious reason I'm
>> missing? Apart from "we are a private company and define our own
>> formats".
>> - If I want to really understand spatial formats used in Oracle
>> Spatial, PostGIS... are the [2] standards the best source? I think
>> they are THE source... Am I right?
>>
>> Thanks in advance, and best regards
>>
>>
>
> --
> Martin Davis
> Senior Technical Architect
> Refractions Research, Inc.
> (250) 383-3022
>
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>